Please note that this is an author-produced PDF of an article accepted for publication following peer review. The definitive publisher-authenticated version is available on the publisher Web site
Journal of Experimental Marine Biology and
Ecology
March 2014, Volume 452, Pages 54–62
Archimer
http://archimer.ifremer.fr
http://dx.doi.org/10.1016/j.jembe.2013.11.014
© 2013 Elsevier B.V. All rights reserved
Contrasting perception of fish trophic level from stomach content and
stable isotope analyses: A Mediterranean artificial reef experience
1, *
Pierre Cresson , Sandrine Ruitton, Mélanie Ourgaud, Mireille Harmelin-Vivien
Aix Marseille Univ, CNRS INSU, IRD, MIO,UM 110, F-13288 Marseille, France.
1
Present address : IFREMER, Lab Environm & Ressources Provence Alpes, Côte d'Azur, Zone portuaire de
Brégaillon, BP 330, 83507 La Seyne sur Mer Cedex
*: Corresponding author : Pierre Cresson, tel.: + 33 494 304 874 ; fax: + 33 494 304 417 ;
email address : [email protected]
Abstract:
A large complex of artificial reefs was deployed in the Bay of Marseilles, North-Western
Mediterranean, for the enhancement of commercial fisheries stocks. Carbon and nitrogen stable
isotope and stomach content analyses were performed on 23 fish species collected on the artificial
reefs to assess their trophic position and feeding behaviour. Results indicated that fish diets were not
modified on the artificial reefs compared to natural environments, nor was the structure of their trophic
network. Artificial reefs, with their complex design, provide diverse and abundant food sources for
13
15
fishes. Ranges of δ C and δ N of artificial reef fishes were comparable to those recorded in natural
Mediterranean environments, with a similar trophic organization. However, some discrepancies
appeared when comparing fish trophic level based on isotopic or diet results, which calls for a careful
interpretation of stable isotope values as direct indicators of trophic level.
Highlights
► Fish diets were assessed on artificial reefs. ► Stomach content and stable isotope data are similar
to those of natural environments. ► Artificial reefs provide diverse food sources for fishes, particularly
15
crustaceans. ► Discrepancy may occur between δ N and fish diet. ► Trophic level estimation should
15
not be based on δ N value alone.
Keywords: Artificial reefs ; Fish diet ; NW Mediterranean ; Stable isotopes ; Stomach content ;
Trophic level
1
1. Introduction
In a context of multiple human impacts on marine ecosystems, artificial reefs have been
widely deployed in marine coastal waters to restore degraded habitats, enhance commercial
and recreational fisheries, promote biodiversity and to protect benthic habitats, among other
management goals (Jensen, 2002; Seaman, 2007). As artificial reefs represent large-scale
experiments, they also provide a way to study ecosystem functioning and to elucidate
ecological processes (Miller, 2002). Ecological hypotheses can be tested by comparing
biological variables (eg. colonization kinetics, recruitment, biomasses, species richness, etc.)
observed on similar modules modified or untreated to serve as controls (Charbonnel et al.,
2002). By the creation of new habitats and increasing food resources, the deployment of
artificial reefs can be a useful tool for enhancing fish biomass and sustaining small-scale
coastal fisheries (Charbonnel et al., 2002; Scarcella et al., 2011; Steimle and Ogren, 1982).
In the Mediterranean Sea, fishing pressures are strong on populations, since 50 % of the
assessed stocks, like mullets or seabreams, are considered overexploited (FAO, 2012).
Artificial reefs are nowadays considered by all stakeholders as an efficient tool to support
small-scale fisheries and also to restore coastal zone under strong fishing pressure (Claudet
and Pelletier, 2004). As fishing targets generally high trophic level species, overfishing is
commonly acknowledged to modify the structure of fish community and decrease fish mean
trophic levels (Pauly et al., 1998). A good understanding of human pressures on marine
ecosystems requires thus robust indicators of fish trophic level.
Studies on trophic patterns of Mediterranean fish have classically and extensively been
performed by stomach content analysis (Bell and Harmelin-Vivien, 1983; Morte et al., 2001;
Rosecchi, 1987; Šantić et al., 2011; and references in Stergiou and Karpouzi, 2002). This
technique allows for the identification of the prey actually consumed by a fish and gives a
“snapshot” of its recent diet. However, some biases linked with accurate prey identification or
different rates of prey digestion may be problematic when using this technique. Moreover,
the low temporal resolution of this technique requires a large number of samples to obtain a
representative view of the dietary patterns of a species (Hyslop, 1980). Some of these biases
can be solved using stable isotope analysis. Isotopic ratios of carbon and nitrogen have been
used to describe trophic relationships in marine Mediterranean ecosystems (Deudero et al.,
2004; Jennings et al., 1997; Pinnegar and Polunin, 2000). When consuming a prey, a
predator integrates the C and N isotopic ratios of its prey into its own tissues. A fractionation
process occurs at each trophic level, as the 13C of the predator is generally slightly higher
than the 13C of its prey (~ + 1 ‰ per trophic level), allowing the use of the carbon isotopic
ratio as an indicator of the organic matter origin. The fractionation factor is higher for nitrogen
(theoretically + 3.4 ‰ per trophic level) and δ15N was classically used as a direct indicator of
the trophic level of the predator (Post, 2002). Nevertheless, due to biases linked with isotopic
ratios of the trophic baseline (ie δ15N value of the source of organic matter at the base of the
trophic network) and variability of nitrogen fractionation factor, some recently published
papers call for a cautious use of δ15N as a direct indicator of trophic level (Mancinelli et al.,
2013; Post, 2002).
Through the “RECIFS PRADO” program, 400 artificial reefs were installed in a 220 ha area
between 25 and 35 m depth in the Bay of Marseilles in 2007 and 2008. “RECIFS PRADO” is
the largest artificial reef program in the Mediterranean Sea and represents the deployment of
a total volume of ~ 27 000 m3 of artificial concrete structures. Its aim is to enhance fish
biomass in the surroundings of artificial reefs and consequently sustain local small-scale
coastal fisheries. This program represented a valuable opportunity 1) to assess the trophic
organization of an artificial reef fish community using stable isotope and stomach content
analyses and 2) to compare the use of δ15N values and diet composition in determining the
trophic level of coastal fish.
2
2. Materials and Methods
Fish were collected on two artificial reefs in the “RECIFS PRADO” zone in the Bay of
Marseilles, France (Fig. 1). These large reefs (6 m high, 187 m 3) are composed of steel and
concrete modules. Their complexity was increased by the addition of bags filled with dead
oyster shells (hereafter named oyster bags) creating shelters for small organisms
(Charbonnel et al., 2011). Their size and complexity provide habitats of different sizes
suitable for most coastal organisms and allow efficient and standardized sampling
procedures. The two artificial reefs investigated, one in the north (V3 reef) and the other in
the south (V6 reef) of the zone, were chosen according to differences in distance from some
organic matter sources (Huveaune River and Posidonia oceanica meadows) and
management status (Cresson et al., 2012). The whole artificial reef zone is currently a full notake area but the southern part will be opened to small-scale artisanal fisheries in a few
years. V3 and V6 artificial reefs are located at 30 m depth on similar sandy bottom with dead
matte of P. oceanica (underlying structure of P. oceanica meadows constituted of rhizomes
and roots intermingled with sediments).
A total of 339 fishes belonging to 32 species were sampled on the artificial reefs by spear
fishing and trammel nets in summer and winter 2010. Species for which only one or two
individuals were collected were discarded and the resulting 325 fishes belonging to 23
species (Table 1) were used for isotopic and stomach content analyses. Details on the
number of fish actually sampled at each season on each reef are presented in Table S1.
In the laboratory, each fish was measured (standard length to the nearest cm) and weighed
(total mass to the nearest g) before dissection. White dorsal muscle was taken for isotopic
analyses before freeze-drying and grinding. In temperate fishes, lipid concentration in white
muscle is generally low and this tissue was demonstrated to be the most suitable for stable
isotope analysis (Pinnegar and Polunin, 1999). Lipid content was assessed by the C/N ratio.
A C/N value lower than 3.5 generaly indicates a lipid content too low to bias the isotopic
ratios (Sweeting et al., 2006). In the whole dataset, less than 10 values were higher than this
threshold and were removed to prevent this bias.
Stable isotope measurements were performed with a continuous-flow isotope-ratio mass
spectrometer (Delta V Advantage, Thermo Scientific, Bremmen, Germany). Results are
expressed in  notation relative to PeeDee Belemnite and atmospheric N2 for 13C and 15N,
 Rsample

 1  10 3 , where X is
respectively, according to the equation X  
R
 standard 
13
C or
15
N and R
is the isotope ratio 13C/12C or 15N/14N, respectively. For both δ13C and δ15N, measurement
precision is < 0.1‰ (replicate measurements of internal laboratory standards, acetanilide).
The trophic level of fish species based on isotopic analysis was assessed using the formula
adapted from Badalamenti et al. (2002): TLi =1+ (δ15Ni – δ15NTB)/3.4, where i is the fish
species, δ15Ni is the nitrogen isotopic ratio for species i, 3.4 the theoretical enrichment at
each trophic level and δ15NTB the nitrogen isotopic ratio for pelagic or benthic primary
production at the base of the trophic network. For pelagic production, the δ15N of
nanophytoplankton (δ15N = 1.77 ‰, Rau et al., 1990) was used as previous results confirmed
its dominance in the Bay of Marseilles (Gregori et al., 2001). Value used for benthic
production (δ15N = 3.91 ‰) is the mean annual value measured for macroalgae sampled on
the artificial reefs (P. Cresson, unpubl. data).
Fish stomachs were removed and stored in 95 % ethanol. Prey items in stomach contents
were sorted under a binocular microscope into their lowest possible taxonomic groups and
their wet weight was obtained to the nearest 0.01 mg. The relative importance of prey taxa in
a fish species‟ diet was assessed by the weight percentage of a food type relative to the total
3
weight of all food ingested. To assess the relative importance of the different prey types for
the whole fish assemblage collected on the artificial reefs, relative importance by weight was
calculated for each prey species. As pisicivores were heavier than other species and
consumed heavy prey, the importance of prey type by weight was corrected by the mean fish
weight. Thus, for each fish species, the consumed mass of each food item was divided by
the mean weight of the species. The cumulative weight of each food item was then
calculated for all fish species consuming this prey. Similarly, the overall occurrence of each
prey type was considered as the percentage of species containing this prey (Hyslop, 1980).
The trophic level of each fish species was issued from bibliographic data based on stomach
content analyses (Barreiros et al., 2002; Darnaude, 2005; Rogdakis et al., 2010; Stergiou
and Karpouzi, 2002; Soares et al., 2003).
Hierarchical clustering based on normalized Euclidean distance and Ward‟s criterion was
performed first on mean isotopic ratios to identify groups of species with similar isotopic
ratios. The same procedure was applied independently to mean stomach content results to
group together species having similar feeding strategies. The results of the two clusterings
were compared. All statistical analyses were performed using R software and the “cluster”
package (Maechler et al., 2012; R Core Team, 2012)
3. Results
3.1. Stable isotope ratios of fishes
Mean isotope values measured for fishes collected on the artificial reefs displayed a 2 ‰
range for 13C (-19.73 to -17.66 ‰) and a 7 ‰ (7.83 to 14.87 ‰) for 15N (Fig. 2), with few
spatial or seasonal differences of fish isotopic ratios (Tab. S2). Six groups of species were
individualized by hierarchical clustering based on their isotopic ratios (SI1 to SI6). SI1 group
comprised three species (Boops boops, Spicara maena and Spicara smaris) and was
characterized by the lowest values of both 13C and 15N (Table 1). Three groups (SI2 to SI4)
with intermediate δ13C and δ15N values were distinguished. One (SI2) was composed of
species belonging to the family Labridae (Coris julis, Symphodus mediterraneus and
Symphodus tinca), and exhibited relatively low values for δ13C and δ15N. The three Diplodus
species (Diplodus vulgaris, D. sargus and D. annularis) clustered together in the SI3 group,
and exhibited relatively high 15N values. The SI4 group, comprising 11 species, was
heterogeneous and displayed high 13C (from - 18.36 to -17.66 ‰) and intermediate δ15N
values (from 9.76 to 10.73 ‰). Sphyraena viridensis clustered apart in the SI5 group with low
δ13C but rather high δ15N (> 11‰). Finally, two species, Dicentrarchus labrax and Trachurus
mediterraneus, formed the SI6 group, which was characterized by the highest δ15N ratios
(> 13‰).
3.2. Stomach content analysis
The analysis of stomach contents revealed the overall importance of crustaceans and fishes
as prey for the whole fish assemblage collected on the artificial reefs (Fig. 3). Crustaceans
were the most frequent prey item consumed (80 % of occurrence), followed by molluscs
(61 %) and polychaetes (52 %). Crustaceans were also the most important prey by weight,
followed by primary producers (macroalgae and P. oceanica) and fishes. Five feeding groups
were identified by hierarchical clustering on stomach contents (Table 2, Fig. 4): “zooplankton
feeders”, “muddy/sand bottom mesocarnivores” (hereafter called soft bottom
mesocarnivores), “rocky/seagrass bed bottom mesocarnivores” (hereafter called rocky
bottom mesocarnivores), “macrocarnivores” and “piscivores”. Feeding group designation took
4
into account not only differences in prey consumed, but prey size (larger in macro- than in
mesocarnivores) and habitat (water column, soft and hard bottom).
3.3. Comparison of isotopic and dietary results
Full concordance between isotopic and feeding groups occurred only for zooplankton feeders
(Fig. 4). The three species of the SI1 group (B. boops, S. maena and S. smaris) fed mainly
on zooplanktonic crustaceans (particularly copepods), even if slight differences could be
observed between them. The diets of S. maena and S. smaris were almost exclusively
composed of zooplanktonic crustaceans (> 75 %), whereas B. boops displayed a more
diverse diet, including fish eggs, gastropods, macroalgae and P. oceanica in addition to
zooplanktonic crustaceans. Rocky bottom mesocarnivores belonged to two different isotopic
groups, SI2 for labrids and SI3 for Diplodus spp. Labrid species consumed mainly bivalves
and polychaetes, but in different proportions (Table 2). S. tinca consumed mostly
polychaetes and to a lesser extent bivalves and gastropods, while bivalves dominated the
diets of C. julis and S. mediterraneus (86 % and 45 % respectively). The diets of the three
Diplodus species (SI3) were more diverse and sessile organisms (macroalgae, ascidians,
bryozoans, cnidarians and hydrozoans) were largely consumed. However, each species
appeared to rely mainly on one food item, primary producers for D. annularis, polychaetes for
D. sargus and, to a lesser extent, ascidians for D. vulgaris. Soft bottom mesocarnivores all
belonged to the SI4 isotopic group. Stomach contents of M. variegatus, M. surmuletus,
P. acarne, P. erythrinus and T. lastoviza were mainly composed of echinoderms and small
benthic crustaceans, along with additional prey like polychaetes or cephalopods (Fig. 5). All
macrocarnivores also belong to the SI4 group. The main prey observed in S. notata, S.
porcus and S. cabrilla were large benthic crustaceans (brachyurans for S. notata and
S. porcus, carids for S. cabrilla, Table 2). The highest discrepancy between isotopic and
feeding clustering occurred for piscivores which were part of three different isotopic groups
(SI4, SI5 and SI6). Synodus saurus, Scorpaena scrofa and Phycis phycis (SI4), Sphyraena
viridensis (SI5), and Dicentrarchus labrax and Trachurus mediterraneus (SI6) displayed
largely different isotopic signatures, while all preyed mainly on fishes (> 85 %, Table 2, Fig.
5)
3.4. Trophic level estimation
The trophic level calculated from stable isotope values using a pelagic baseline matched the
trophic level estimation based on stomach contents for most trophic groups: zooplankton
feeders, labrids, soft bottom mesocarnivores, macrocarnivores and pelagic piscivores (Table
3). Calculations based on benthic isotopic baseline always indicated lower trophic levels for
these groups. On the contrary, the trophic level of Diplodus spp from stomach contents better
fitted the isotopic calculation based on benthic baseline, while the pelagic baseline led to a
higher estimation. Eventually, the trophic level of benthic piscivores was always lower when
calculated from isotopic values than from stomach contents, whatever the trophic baseline
used.
4. Discussion
With the exception of gobiids, blennids and pomacentrids which could not be sampled due to
their small size, the present study analysed the majority of the fish species observed by
underwater visual censuses on the artificial reefs in the Bay of Marseilles (Rouanet et al.,
2012). The most abundant species (sparids, scorpaenids, serranids, mullids) included in our
5
study dominated also in natural rocky habitats (Fasola et al., 1997; Harmelin, 1987;
Letourneur et al., 2003).
4.1. Use of artificial reef food resources by fishes
Fish diets on the artificial reefs were rather similar with those observed in natural
environments in the area of Marseilles (Bautista-Vega et al., 2008; Bell and Harmelin-Vivien,
1983; Harmelin-Vivien et al., 1989) or in other Mediterranean zones (Fanelli et al., 2011;
Kalogirou et al. 2012; Quignard, 1966; Sala and Ballesteros, 1997; Stergiou and Karpouzi,
2002), even if no direct comparison of diets in natural and artificial habitats could be
performed in this work.
Stomach contents observed for Boops boops, Spicara smaris and S. maena placed them in
the cluster of secondary consumers specialized on zooplanktonic organisms, as already
observed (Bell and Harmelin-Vivien, 1983; Stergiou and Karpouzi, 2002). Different
proportions of zooplankton and vegetal material were observed in the diet of the more
opportunistic B. boops (Derbal and Kara, 2008; Fasola et al., 1997).
The majority of the fish species sampled on the artificial reefs exhibited a mesocarnivorous
diet, based on small invertebrates (crustaceans, molluscs, echinoderms or polychaetes) and
benthic primary producers. Taking into consideration the habitat of species allowed clearly
distinguishing between rocky and sandy bottom fishes. Based on their diets, labrids and
Diplodus species were gathered in the same feeding cluster of rocky bottom mesocarnivores.
Labrids mainly consumed small benthic invertebrates like molluscs and polychaetes. C. julis
is known to prey on gastropods and crustaceans (Quignard, 1966), juvenile echinoderms
(Sala, 1997) or bivalves (Bell and Harmelin-Vivien, 1983). The diets of S. tinca and
S. mediterraneus are similar and composed of small crustaceans, bivalves, gastropods and
polychaetes (Bell and Harmelin-Vivien, 1983). Diplodus spp. can be considered as
omnivores as they presented a diversified diet, feeding on a large range of prey, from
primary producers to fishes, with a high consumption of diverse sessile invertebrates.
Omnivory of Diplodus species is well documented (Derbal et al., 2007; Rosecchi, 1987; Sala
and Ballesteros, 1997). The soft bottom mesocarnivores (M. variegatus, M. surmuletus, P.
acarne, P. erythinus and T. lastoviza) preyed mainly on echinoderms and small crustaceans,
along with other diverse prey like molluscs or polychaetes, consistently with previous results
(Bautista-Vega et al., 2008; Fanelli et al., 2011; Fehri-Bedoui et al., 2009). The diet of
macrocarnivores (S. notata, S. porcus and S. cabrilla) was also composed of crustaceans,
but of larger size (brachyurans and carids), and in larger proportion (> 90 % of the ingested
prey weight) than mesocarnivores (Harmelin-Vivien et al., 1989). For all these species, the
present study highlighted the importance of crustaceans in artificial reef functioning, which
represented the most commonly ingested prey type. These results confirm the major role of
crustaceans in fish diets on artificial reefs (Leitão et al., 2007; Relini et al., 2002). Finally, six
species (Scorpaena scrofa, Phycis phycis, Trachurus mediterraneus, Sphyraena viridensis,
Synodus saurus and Dicentrarchus labrax) mainly preyed on fish and could be considered as
piscivores. Even if all these predators consume fishes, differences appeared nevertheless
amongst them, as they occupied different habitats and consumed different species. Previous
works report a high consumption of zooplankton feeding species (B. boops and S. smaris) by
S. viridensis, which could be consider as a pelagic piscivore (Kalogirou et al., 2012). P.
phycis could be consider as a benthic piscivore, as remains of benthic species (C. julis and
S. tinca) were observed in its stomach contents on the artificial reefs. Previous works
indicate also the consumption of B. boops, Spicara spp. or Chromis chromis by S. saurus
and S. scrofa in natural environments (Esposito et al., 2009; Šantić et al., 2011). Finally,
T. mediterraneus and D. labrax were high trophic level transient piscivores, as confirmed by
the predominance of fish remains in their stomach, consistently to previous works (Pasquaud
et al., 2010; Rogdakis et al., 2010).
6
4.2. Stable isotopes and the trophic structure of fish community
Fishes sampled on the artificial reefs of Marseilles displayed a range of isotopic ratios (~ 2 ‰
for 13C and 7 ‰ for 15N) similar to those observed in previous studies of fish assemblages
in natural Mediterranean coastal rocky environments (Jennings et al., 1997; Pinnegar and
Polunin, 2000; Vizzini and Mazzola, 2009). Using δ15N as an indicator of trophic level, it could
be considered that fishes occupied at least three trophic levels on the artificial reefs studied.
The lower level (δ15N < 9 ‰) was occupied by the three zooplankton-feeders (B. boops, S.
smaris and S. maena) which clustered in the SI1 group, the only one to be conserved
between the two classifications. The specificity of their diets, both on specific and isotopic
points of view, could explain the robustness of this group. The second intermediate trophic
level (9 ‰ < δ15N < 13 ‰) gathered most of the species sampled on the artificial reefs.
However, four distinct isotopic groups of intermediate trophic level fishes were separated
based on both their δ15N and δ13C values, and these groups did not match with stomach
content clustering. Within the rocky bottom mesocarnivorous group, difference in isotopic
values placed labrids (SI2) and Diplodus spp. (SI3) in two distinct isotopic groups. The rather
low isotopic values observed for labrids were consistent with the consumption of small low
trophic level invertebrates, as previously observed in natural environments (Bell and
Harmelin-Vivien, 1983; Quignard, 1966; Sala, 1997). Contrarily, the δ15N observed for
Diplodus spp. was surprisingly high for species consuming small invertebrates and primary
producers, as it was higher than those of carnivorous species like S. porcus. A similar pattern
between δ15N ratios measured in Diplodus spp. and S. porcus is reported in other
Mediterranean rocky environments. In Formentera, Balearic Islands, D. annularis displays a
higher δ15N ratio than S. porcus (9.44 ‰ and 8.93 ‰ respectively, Deudero et al., 2004). In
the Bay of Calvi, Corsica, Pinnegar and Polunin (2000) report also higher δ15N values for
D. annularis and D. sargus than for S. porcus (8.39, 9.13 and 7.93 ‰ respectively). These
authors explain such high δ15N values by the consumption of fishes by Diplodus spp., but
have not analysed fish stomach contents. Such an explanation is likely not pertinent here
due to the minor importance of fish in the diet of Diplodus spp. collected on the artificial reefs,
and is confirmed by other studies on their feeding habits (Bell and Harmelin-Vivien, 1983;
Derbal et al., 2007; Sala and Ballesteros, 1997). The apparent discrepancy between δ15N
values and diet for Diplodus spp. calls for cautious interpretation of isotopic ratios (see last
paragraph). The SI4 group was the most diverse of intermediate trophic level fishes, as it
gathered species classified with stomach contents as soft-bottom mesocarnivores,
macrocarnivores and benthic piscivores. Finally, the pelagic Sphyraena viridensis clustered
apart (SI5) at an intermediate trophic level in spite of a piscivorous diet. The highest trophic
level (δ15N > 13 ‰) was represented by two well-known piscivores, Dicentrarchus labrax and
Trachurus mediterraneus (Rogdakis et al., 2010; Stergiou and Karpouzi, 2002).
Thus, in spite of apparently similar diets based on fish, piscivorous species displayed
different isotopic ratios and were classified in three isotopic groups (SI4, SI5 and SI6). The
comparable δ13C values of Sphyraena viridensis (SI5) and Boops boops, Spicara maena and
S. smaris, along with the ~ 3 ‰ higher δ15N of S. viridensis, testified of the consumption of
zooplanktivorous fishes by this piscivore, as observed by Kalogirou et al. (2012). The three
benthic piscivorous species (S. saurus, S. scrofa and P. phycis) displayed δ15N values close
to those of crustacean-eating mesocarnivores (SI4). This might appear surprising for fish
consumers, generally considered to be at the highest trophic level in the Mediterranean
(Stergiou and Karpouzi, 2002), but could be explained by the consumption of low trophic
level benthic fish species. Finally, the high 15N value of T. mediterraneus and D. labrax (SI6)
confirmed the high trophic level of these two piscivores (Pasquaud et al., 2010; Rogdakis et
al., 2010).
7
4.3. δ15N: a direct indicator of trophic level?
Carbon and nitrogen stable isotope ratios have been widely used to assess trophic patterns
in fish communities, sometimes replacing time-consuming stomach content analysis
(Deudero et al., 2004; Jennings et al., 1997; Pinnegar and Polunin, 2000). In this approach
δ15N is considered as a direct indicator of the trophic position of consumers. However,
recently published results demonstrated that different factors, like trophic baseline, isotopic
fractionation and consumer metabolism among others, may bias direct trophic level
interpretation (Mancinelli et al., 2013; Mill et al., 2007; Post 2002; Schmidt et al., 2004).
Results of the current study illustrated some of these biases. Discrepancies were observed
between the classifications based on stable isotope or stomach content analyses for all
trophic groups except zooplankton feeders. The case of Diplodus spp. illustrated the
influence of both trophic baseline isotopic ratios and trophic enrichment factor on the trophic
level estimation of herbivorous (or omnivorous) versus carnivorous species (Mancinelli et al.,
2013; Mill et al., 2007). A simplistic interpretation of the high Diplodus δ15N could place these
species at a high trophic level, just below the piscivorous D. labrax and T. mediterraneus, as
suggested by the calculation of their trophic level from isotopic values. Such a position was
incongruous with their omnivorous feeding patterns, as they consumed large amounts of
algal matter and sessile invertebrates, along with other low trophic level prey. Diplodus δ15N
(~ 11 ‰) was surprisingly ~ 2 ‰ higher than those of labrids, which belonged to the same
trophic group of rocky bottom mesocarnivores (Fig. 4). Their δ15N value could reflect the
importance of algae or seagrass in their diet as benthic primary producers classically present
higher δ15N than phytoplankton (Nadon and Himmelman, 2006). The calculation of Diplodus
spp. trophic level using a benthic baseline, rather than a pelagic one, resulted actually in a
more consistent trophic level value, similar to the labrids one, and lower than those of
piscivores. Recent works demonstrated also that fractionation factor associated with
herbivory is higher than the usually-accepted 3.4 ‰ value, due to differences in the
enzymatic and digestive systems of herbivorous species (Mill et al., 2007; Wyatt et al., 2010).
The combined effects of prevalent benthic vegetal material in the diet of Diplodus spp., along
with a higher fractionation factor, explained thus likely their high δ15N values. This
demonstrated how direct simplistic interpretation of isotopic ratios can lead to wrong
inferences of consumers‟ diet and trophic level.
In the Mediterranean Sea, piscivorous species are considered to represent the highest
trophic level for fishes, with a mean value close to or higher than 4 (Stergiou and Karpouzi,
2002). In the present study, the large range of δ15N values (10.1 - 14.9 ‰) recorded in the six
piscivorous species analysed would place them at two distinct trophic levels. The lower
trophic level of S. scrofa and P. phycis, when calculated from stable isotopes, was probably
due to their consumption of crustaceans, in addition to fish. But some strictly piscivorous
species, like S. viridensis and S. saurus, displayed an isotopic-calculated trophic level lower
than 4, as they consume zooplankton-eating or small benthic fishes, which presented
themselves low δ15N values. Inferring their diet and trophic level only from their δ15N values
that were similar to those of meso- and macrocarnivores (Fig.5), would have led to
misinterpretations. On the contrary, the high δ15N of D. labrax and T. mediterraneus were
consistent with a trophic level higher than 4, as previously calculated by Rogdakis et al.
(2010). Using the specific 3.8 ‰ fractionation factor proposed for fish muscle δ15N by
Sweeting et al. (2007), D. labrax prey should exhibit δ15N values close to 10 ‰, similar to
those recorded for macrocarnivorous and other piscivorous fish species, which confirmed the
consumption of high trophic level fishes by this predator. The current results confirmed the
separation amongst fish-eating species proposed by Stergiou and Karpouzi (2002). A trophic
level lower than 4 appear consistent for species eating fish and crustaceans. But a trophic
level close to or higher than 4 for all fish-eating species can be questioned, since some
species exhibit low δ15N while preying only on fishes. Thus, “piscivore” is not always
synonymous with “top-predator”.
8
The results obtained on the feeding habits and trophic position of artificial reef fishes in the
Bay of Marseilles demonstrates that combining stable isotope and stomach content analyses
remains a necessary approach to clarify fish trophic relationships, especially given that
different prey types may present similar isotopic signatures (Layman et al., 2012). Even
within one ecosystem, the crude comparison of δ15N values of organisms does not
necessarily reflect their trophic level and diet composition. The isotopic values measured in a
consumer result not only from the isotopic ratios of its prey, but also from varying
fractionation factors and food component routine (Perga and Grey, 2010; Schmidt et al.,
2004; Sweeting et al., 2007). Thus, the correct interpretation of trophic relationships of
organisms based on stable isotopic signatures is a highly complex task, which does not
preclude the knowledge of their biology and feeding behaviour.
Acknowledgements
All fishes sampled with trammel nets were taken during experimental fishing by J.-Y.
Jouvenel and members of “P2A Développement” team to whom we are grateful. Thanks are
also expressed to B. de Ligondes F. Zuberer, G. Bleton (Scuba diving crew, OSU Pytheas)
and to F. Morat for their most valuable help during field sampling, especially in winter. Stable
isotope analyses were performed by P. Richard and G. Guillou (LIENSs laboratory,
Université de la Rochelle). We thank also the “GIS POSIDONIE” team for providing
underwater censuses results and for valuable discussions and comments. English
corrections of a previous version of the manuscript were done by Rachel Mackie.
This work is part of the „„RECIFS PRADO‟‟ program and was funded by grants from the city
of Marseilles and the “Agence de l‟Eau Rhône-Mediterranée-Corse”.
References
Badalamenti, F., D‟Anna, G., Polunin, N.V.C., 2002 Size-related trophodynamic changes in
three target fish species recovering from intensive trawling. Mar. Biol. 141: 561-570.
Barreiros, J. P., Santos, R. S., de Borba, A. E., 2002. Food habits, schooling and predatory
behaviour of the yellowmouth barracuda, Sphyraena viridensis (Perciformes:
Sphyraenidae) in the Azores. Cybium 26, 83-88.
Bautista-Vega, A.A., Letourneur, Y., Harmelin-Vivien, M., Salen-Picard, C., 2008. Difference
in diet and size-related trophic level in two sympatric fish species, the red mullets
Mullus barbatus and Mullus surmuletus, in the Gulf of Lions (north-west
Mediterranean Sea). J. Fish Biol. 73, 2402-2420.
Bell, J.D., Harmelin-Vivien, M.L., 1983. Fish fauna of French Mediterranean Posidonia
oceanica seagrass meadow. 2. Feeding habits. Tethys 11, 1-14.
Charbonnel, E., Serre, E., Ruitton, S., Harmelin, J.-G., Jensen, A.C., 2002. Effects of
increased habitat complexity on fish assemblages associated with large artificial reef
units (French Mediterranean coast). ICES J. Mar. Sci. 59, S208–S213.
Charbonnel, E., Harmelin, J.G., Carnus, F., Le Direac'h, L., Ruitton, S., Lenfant, P., Beurois,
J., 2011. Artificial reefs in Marseille (France, Mediterranean Sea): From complex
natural habitat to concept of efficient artificial reef design. Braz. J. Oceanogr. 59, 177178.
Claudet, J., Pelletier, D., 2004. Marine Protected areas and artificial reefs: A review of the
interactions between management and scientific studies. Aq. Liv. Ress. 17, 129-138.
9
Cresson, P., Ruitton, S., Fontaine, M.-F., Harmelin-Vivien, M., 2012. Spatio-temporal
variation of suspended and sedimentary organic matter quality in the Bay of
Marseilles (NW Mediterranean) assessed by biochemical and isotopic analyses. Mar.
Pollut. Bull. 64, 1112-1121.
Darnaude, A. M., 2005. Fish ecology and terrestrial carbon use in coastal areas: implications
for marine fish production. J. Anim. Ecol. 74, 864-876.
Derbal, F., Kara, M.H., 2008. Composition du régime alimentaire du bogue Boops boops
(Sparidae) dans le golfe d'Annaba (Algérie). Cybium 32, 325-333.
Derbal, F., Nouacer, S., Kara, M.H., 2007. Composition et variation du régime alimentaire du
sparaillon Diplodus annularis (Sparidae) du golfe d'Annaba (Algérie). Cybium 32,
325-333.
Deudero, S., Pinnegar, J.K., Polunin, N.V.C., Morey, G., Morales-Nin, B., 2004. Spatial
variation and ontogenic shifts in the isotopic composition of Mediterranean littoral
fishes. Mar. Biol. 145, 971-981.
Esposito, V., Battaglia, P., Castriota, L., Finoia, M.G., Scotti, G., Andaloro, F., 2009. Diet of
atlantic lizardfish Synodus saurus (Linnaeus, 1759) (Pisces: Synodontidae) in the
Central Mediterranea Sea. Sci. Mar. 78, 369-376.
Fanelli, E., Baldamenti, F., D'Anna, G., Pipitone, C., Riginella, E., Azzuro, E., 2011. Food
partitioning and diet temporal variation in two coexisting sparids Pagellus erythinus
and Pagellus acarne. J. Fish Biol. 78, 869-900.
FAO, 2012. The state of world fisheries and aquaculture 2012, Rome, 209 pp.
Fasola, M., Canova, L., Foschi, F., Novelli, O., Bressan, M., 1997. Resource Use by a
Mediterranean Rocky Slope Fish Assemblage. P. Z. N. I: Marine Ecology 18, 51-66.
Fehri-Bedoui, R., Mokrani, E., Ben Hassine, O.K., 2009. Feeding habits of Pagellus acarne
(Sparidae) in the Gulf of Tunis, central Mediterranean. Sci. Mar. 73, 667-678.
Gregori, G., Colosimo, A., Denis, M., 2001. Phytoplankton group dynamics in the Bay of
Marseilles during a 2-year survey based on analytical flow cytometry. Cytometry. 44,
247-256.
Harmelin, J.-G., 1987. Structure et variabilité de I'ichtyofaune d'une zone rocheuse protégée
en Méditerranée (Parc national de Port-Cros, France). P. Z. N. I: Mar. Ecol. 8, 263284.
Harmelin-Vivien, M.L., Kaim-Malka, R.A., Ledoyer, M., Jacob-Abraham, S.S., 1989. Food
partitionning among scorpaenid fishes in Mediterranean seagrass bed. J. Fish Biol.
34, 715-734.
Hyslop, E.J., 1980. Stomach contents analysis-a review of methods and their application. J.
Fish Biol. 17, 411-129.
Jennings, S., Reñones, O., Morales-Nin, B., Polunin, N.V.C., Moranta, J., Coll, J., 1997.
Spatial variation of the 15N and 13C isotopes composition of plant, invertebrates and
fishes, on Mediterranean reefs: implications for the study of trophic pathways. Mar.
Ecol. Progr. Ser. 146, 109-116.
Jensen, A.C., 2002. Artificial reefs of Europe: perspective and future. ICES J. Mar. Sci. 59,
S3-S13.
Kalogirou, S., Mittermayer, F., Pihl, L., Wennhage, H., 2012. Feeding ecology of indigenous
and non-indigenous fish species within the family Sphyraenidae. J. Fish Biol. 80,
2528-2548.
Layman, C.A., Araujo, M.S., Boucek, R., Hammerschlag-Peyer, C.M., Harrison, E., Jud, Z.R.,
Matich, P., Rosenblatt, A.E., Vaudo, J.J., Yeager, L.A., et al., 2012. Applying stable
isotopes to examine food-web structure: an overview of analytical tools. Biol. Rev. 87,
545-562.
Leitão, F., Santos, M.N., Monteiro, C.C., 2007. Contribution of artificial reefs to the diet of the
white sea bream (Diplodus sargus). ICES J. Mar. Sci. 64, 473-478.
Letourneur, Y., Ruitton, S., Sartoretto, S., 2003. Environmental and benthic habitat factors
structuring the spatial distribution of a summer infralittoral fish assemblage in the
North-Western Mediterranean Sea. J. Mar. Biol. Ass. UK 83, 193-204.
10
Maechler, M., Rousseeuw, P., Struyf, A., Hubert, M., Hornik, K.,2012. cluster: Cluster
Analysis Basics and Extensions. R package version 1.14.2.
Mancinelli, G., Vizzini, S., Mazzola, A., Maci, S., Basset, A. 2013. Cross-validation of δ15N
and FishBase estimates of fish trophic position in a Mediterranean lagoon: the
importance of isotopic baseline. Est. Coast. Shelf Sci. doi:
10.1016/j.ecss.2013.04.004.
Mill, A.C., Pinnegar, J.K., Polunin, N.V.C., 2007. Explaining isotope trophic-step fractionation:
why herbivorous fish are different. Funct. Ecol. 21, 1137-1145.
Miller, M.W., 2002. Using ecological processes to advance artificial reef goals. ICES J. Mar.
Sci. 59, 27-31.
Morte, S., Redon, M.J., Sanz-Brau, A., 2001. Diet of Scorpaena porcus and Scorpaena
notata (Pisces : Scorpaenidae) in the Western Mediterranean. Cah. Biol. Mar. 42,
333-344.
Nadon, M.-O., Himmelman, J.H., 2006. Stable isotopes in subtidal food webs: Have enriched
carbon ratios in benthic consumers been misinterpreted? Limnol. Oceanogr. 51,
2828-2836.
Pauly, D., Christensen, V., Dalsgaard, J., Froese, R., Torres, F., 1998. Fishing Down Marine
Food Webs. Science 279, 860-863.
Pasquaud, S., Pillet, M., David, V., Sautour, B., Elie, P., 2010. Determination of fish trophic
level in an estuarine system. Estuar. Coast. Shelf Sci. 86, 237-246.
Perga, M.E., Grey, J., 2010. Laboratory measures of isotope discrimination factors:
comments on Caut, Angulo & Courchamp (2008, 2009). J. Appl. Ecol. 47, 942-947.
Pinnegar, J.K., Polunin, N.V.C., 1999. Differential fractionation of 13C and 15N among fish
tissues : implication for the study of trophic interactions. Funct. Ecol. 13, 225-231.
Pinnegar, J.K., Polunin, N.V.C., 2000. Contributions of stable-isotope data to elucidating food
webs of Mediterranean rocky littoral fishes. Oecologia 122, 399-409.
Post, D.M., 2002. Using stable isotopes to estimate trophic position: models, methods, and
assumptions. Ecology 83, 703-718.
Quignard, J.P., 1966. Recherches sur les labridae (poissons téléostéens perciformes) des
côtes européennes. Systématique et biologie. Causse et Castelnau, Montpellier, 1248 pp.
R Core Team (2012). R: A language and environment for statistical computing. R Foundation
for Statistical Computing, Vienna, Austria.
Rau, G.H., Teyssie, J.L., Rassoulzadegan, F., Fowler, S.W., 1990. 13C/12C and 15N/14N
variations among size-fractionated marine particles: implications for their origin and
trophic relationships Mar. Ecol.-Prog. Ser. 59, 33-38.
Relini, G., Relini, M., Torchia, G., De Angelis, G., 2002. Trophic relationship between fishes
and an artificial reef. ICES J. Mar. Sci. 59, S36-S42.
Rogdakis, Y., Ramfos, A., Koukou, K., Dimitriou, E., Katselis, G., 2010. Feeding habits and
trophic level of sea bass (Dicentrarchus labrax) in the Messolonghi-Etoliko lagoons
complex (Western Greece). J. Biol. Res-Thessalon. 13, 13-26.
Rosecchi, E., 1987. L'alimentation de Diplodus annularis, Diplodus sargus, Diplodus vulgaris
et Sparus aurata (Pisces : Sparidae) dans le golfe du Lion et les lagunes littorales.
Rev. Trav. Inst. Pêches Marit. 49, 125-141.
Rouanet, E., Astruch, P., Le Direac'h, L., Bonhomme, D., Bonhomme, P., 2012. Opération
Récifs Prado: Suivi scientifique, biologique et technique dans la zone d'immersion
(suvi obligatoire). Rapport annuel 2012. Résultats. GIS Posidonie - Ville de Marseille,
Fr, pp. 1-264.
Sala, E., 1997. Fish predator and scavengers of the sea urchin Paracentrotus lividus in
protected areas of the North-Western Mediterranean Sea. Mar. Biol. 129, 531-539.
Sala, E., Ballesteros, E., 1997. Partitionning of space and food ressources by three fishes of
the genus Diplodus (Sparidae) in a Mediterranean rocky infralittoral ecosystem. Mar.
Ecol. Progr. Ser. 152, 273-283.
11
Šantić, M., Pallaoro, A., Stagličić, N., Markov-Podvinski, M., 2011. Feeding habits of the red
scorpionfish, Scorpaena scrofa (Osteichthyes: Scorpaenidae) from the eastern
central Adriatic Sea. Cah. Biol. Mar. 52, 217-226.
Scarcella, G., Grati, F., Polidori, P., Domenichetti, F., Bolognini, L., Fabi, G., 2011.
Comparison of growth rates estimated by otolith reading of Scorpaena porcus and
Scorpaena notata caught on artificial and natural reefs of the northern Adriatic Sea.
Braz. J. Oceanogr. 59, 32-42.
Seaman, W., 2007. Artificial habitats and the restoration of degraded marine ecosystems and
fisheries. Hydrobiologia 580, 143-155.
Schmidt, K., McClelland J.W., Mente, E., Montoya, J.P., Atkinson, A., Voss, M., 2004.
Trophic-level interpretation based on δ15N values: implication of tissue-specific
fractionation and amino-acid composition. Mar. Ecol. Prog. Ser. 266, 43-58.
Soares, M.S.C., Sousa, L., Barreiros, J. P., 2003. Feeding habits of the lizardfish Synodus
saurus (Linnaeus, 1758) (Actinopterygii: Synodontidae) from the Azores. Aqua, J.
Ichthyol. Aquat. Biol. 7, 29-38.
Steimle, F.L., Ogren, L., 1982. Food of fish collected on artificial reefs in the New York bight
and Off Charleston, South Carolina. Mar. Fish. Rev. 44, 49-52.
Stergiou, K.I., Karpouzi, V.S., 2002. Feedings habits and trophic levels of Mediterranean
fishes. Rev. Fish. Biol. Fisher. 11, 217-254.
Sweeting, C.J., Barry, J., Barnes, C., Polunin, N.V.C., Jennings, S., 2007. Effects of body
size and environment on diet-tissue δ15N fractionation in fishes. J. Exp. Mar. Biol.
Ecol. 340, 1-10.
Sweeting, C.J., Polunin, N.V.C., Jennings, S., 2006. Effects of chemical lipid extraction and
arithmetic lipid correction on stable isotope ratios of fish tissues. Rapid Commun.
Mass Sp. 20, 595-601.
Vizzini, S., Mazzola, A., 2009. Stable isotopes and trophic position of littoral fishes from a
Mediterranean marine protected area. Environ. Biol. Fishes 84, 13-25.
Wyatt, A.S.J., Waite, A.M., Humphries, S., 2010. Variability in isotope discrimination factors
in coral reef fishes: implications for diet and food web reconstruction. PLoS ONE
5(10), e13682. doi:10.1371/journal.pone.0013682.
12
Figures
Fig. 1: Location of artificial reefs (V3 and V6) in the Bay of Marseilles.
13
Fig. 2: Mean isotopic ratios (13C and 15N) of fishes sampled on artificial reefs. Species
grouped together by the hierarchical clustering analysis are represented with similar symbols
(white square: SI1; circle: SI2; triangles: SI3, diamonds: SI4, black square: SI5; crosses:
SI6). Colours of the diamonds in the SI3 group represent the diet of the species (black:
piscivores; grey: macrocarnivores; white: muddy/ sand bottom mesocarnivores). For
graphical convenience, standard deviations are not plotted and the names of some species
are abbreviated (S. medit: Symphodus mediterraneus, S. not: Scorpaena notata, M. sur:
Mullus surmuletus, P. ery: Pagellus erythrinus).
14
Fig. 3: Percentage of occurrence (black bars) and cumulative corrected mass (grey bars) of
the main taxonomic groups of prey consumed by all the fish species analysed. Cnid. :
Cnidarians
Fig.4: Hierarchical clustering tree based on stomach content analysis. The superimposed SI1
to SI6 indications represent the isotopic groups formed by an independent hierarchical
clustering based on stable isotope ratios. They were added to evidence differences in
clustering run with the two data sets.
15
Fig. 5: Food spectra of nine species belonging to the SI4 group according to similar SI ratios
but to three trophic groups according to different diets. Numbers after species names
represent the number of stomachs analysed.
16
Tables
Table 1: Family, species, mean standard length (SL) and range (minimal and maximal length), mean mass and range (minimal and maximal
mass), stable isotope ratios (δ13C ± sd, δ15N ± sd); numbers of fishes (n indiv), of stable isotope analyses (n SIA) and of stomach content
analyses (SCA) realized.
Family
Species
Carangidae
Trachurus
mediterraneus
(Steindachner,
1868)
Spicara maena
(Linnaeus, 1758)
Spicara smaris
(Linnaeus, 1758)
Coris julis
(Linnaeus, 1758)
Symphodus
mediterraneus
(Linnaeus, 1758)
Symphodus tinca
(Linnaeus, 1758)
Dicentrarchus
labrax
(Linnaeus, 1758)
Mullus surmuletus
Linnaeus, 1758
Phycis phycis
(Linnaeus, 1766)
Scorpaena notata
Rafinesque, 1810
Centracanthidae
Labridae
Moronidae
Mullidae
Phycidae
Scorpaenidae
SL (mm)
Mass (g)
13C (‰)
sd
15N
(‰)
sd
n indiv
n SIA
n SCA
259 (236-277)
231 (215-245)
-18.74
0.77
14.87
2.98
5
10
4
131 (104-170)
62 (35-101)
-19.51
0.33
7.83
0.25
10
26
6
141 (116-154)
63 (45-104)
-19.73
0.41
8.61
0.77
9
13
5
109 (68-144)
19 (5-45)
-18.62
0.55
9.74
0.31
13
18
13
109 (95-131)
41 (25-75)
-19.25
0.26
9.17
0.52
5
18
16
158 (115-213)
110 (35-250)
-18.78
0.84
9.55
0.52
6
15
12
325 (267-430)
811
(370- 1870)
-18.20
2.53
13.92
0.6
4
12
3
142 (91-216)
75 (20-225)
-17.88
0.72
9.94
0.67
35
44
19
334 (333-335)
668 (650-685)
-17.95
0.3
10.73
0.33
3
9
3
112 (68-148)
65 (10-195)
-17.66
0.39
10.15
0.5
33
41
27
17
Serranidae
Soleidae
Sparidae
Sphyraenidae
Synodontidae
Triglidae
Scorpaena porcus
Linnaeus, 1758
Scorpaena scrofa
Linnaeus, 1758
Serranus cabrilla
(Linnaeus, 1758)
Microchirus
variegatus
(Donovan, 1808)
Boops boops
(Linnaeus, 1758)
Diplodus annularis
(Linnaeus, 1758)
Diplodus sargus
(Linnaeus, 1758)
Diplodus vulgaris
(Geoffroy SaintHilaire, 1817)
Pagellus acarne
(Risso, 1827)
Pagellus
erythrinus
(Linnaeus, 1758)
Sphyraena
viridensis
Cuvier, 1829
Synodus saurus
(Linnaeus, 1758)
Trigloporus
lastoviza
(Bonnaterre, 1788)
142 (84-251)
128 (25-400)
-17.67
0.49
9.74
0.41
28
37
6
177 (124-217)
275(80-535)
-18.07
0.22
10.06
0.23
5
18
3
139(114-169
61(25-105)
-18.36
0.22
9.79
0.23
20
43
14
87 (72-100)
19 (10-30)
-18.14
0.34
10.2
0.14
7
21
6
158 (101-222)
69 (30-195)
-19.90
0.29
8.46
0.32
33
33
30
123 (94-188)
66 (25-185)
-18.85
1.1
11.72
1.65
48
34
28
158 (142-195)
158 (100-285)
-18.77
0.69
11.54
0.73
5
13
4
115 (65-173)
76 (10-195)
-18.14
0.78
11.59
0.83
20
58
12
118 (97-190)
47 (20-145)
-17.76
0.6
10.64
0.56
20
36
11
147 (115-162)
72 (40-105)
-17.77
0.26
10.81
0.6
5
32
3
394 (373-414)
355 (295-395)
-19.33
0.85
11.07
0.65
5
10
3
207 (161-235)
115 (40-170)
-18.30
0.46
10.5
0.79
3
9
2
148 (102-195)
76 (20-155)
-17.9
0.56
9.76
1.07
3
9
2
18
Table 2: Weight percentage for prey taxa in stomach content, prey representing less than 1% are presented with + symbol. Trophic group were
determined by hierarchical clustering on stomach contents. Zoo: zooplankton feeders, MesoRB: rocky/ seagrass bottom mesocarnivores,
MesoSB: muddy/ sandy bottom mesocarnivores, Macroc: macrocarnivores, Pisc: piscivores. SI group: groups resulting from the clustering
analysis on SI ratios. Prim prod: primary producers, Cepha: cephalopods, Gastr: gastropods, Bival: bivalves, Polyc: polychaetes, Zoopk:
zooplanktonic crustaceans, Crust: benthic crustaceans, Echin: echinoderms, Ascid: ascidians. Groups with minor contribution were pooled in
“Others” and represented by superscript letters (b: bryozoans, c: cnidarians, h: hydrozoans, f: foraminifera, u: unidentified matter).
Boops boops
Spicara smaris
Spicara maena
Trophic
group
Zoo
Zoo
Zoo
Coris julis
Symphodus tinca
S. mediterraneus
Diplodus annularis
Diplodus sargus
Diplodus vulgaris
MesoRB
MesoRB
MesoRB
MesoRB
MesoRB
MesoRB
2
2
2
3
3
3
3
87
6
-
-
+
10
6
+
+
6
86
21
47
+
1
-
3
41
+
+
86
10
-
11
8
6
+
-
-
-
-
8
2
7
47
6
-
29b+f+u
37u
b+h+f+u
2
29 b+h
Microchirus variegatus
Mullus surmuletus
Pagellus acarne
Pagellus erythrinus
Trigloporus lastoviza
MesoSB
MesoSB
MesoSB
MesoSB
MesoSB
4
4
4
4
4
2
3
-
62
11
+
6
32
-
13
6
-
21
13
7
-
-
19
47
20
45
38
49
24
56
24
-
2
-
-
+u
-
Scorpaena notata
Scorpaena porcus
Macroc
Macroc
4
4
+
-
-
3
-
+
-
97
99
-
-
-
-
Species
SI group
Prim prod
Cepha
Gastr Bival Polyc Zoopk Crust
1
1
1
4
2
-
-
10
20
-
1
-
9
69
78
91
Echin
Ascid
Fish
Others
-
-
-
14
-
2 c+u
-
19
Table 2 (continued)
Serranus cabrilla
Macroc
4
-
-
-
-
+
-
91
-
-
9
-
Synodus saurus
Scorpaena scrofa
Phycis phycis
Sphyraena viridensis
Dicentrarchus labrax
Trachurus mediterraneus
Pisc
Pisc
Pisc
Pisc
Pisc
Pisc
4
4
4
5
6
6
-
-
-
-
-
-
15
4
+
1
-
-
100
85
96
100
100
99
-
20
Table 3: Comparison of trophic levels of artificial reef fishes. Trophic levels (TL) were issued
from bibliographical data when based on stomach contents, with superscript letters standing
for the reference used, or were calculated from the stable isotopic values recorded in this
study, using the formula of Badalamenti et al. (2002): NTi = 1 + (δ15Ni – δ15NB) / 3.4, with
δ15Ni the δ15N value of fish species i, 3.4 the theoretical trophic enrichment factor, and δ 15NB
the isotopic ratio of the trophic baseline. Nanophytoplankton (δ15N = 1.77 ‰) or macroalgae
(δ15N = 3.91 ‰) were used as proxies of pelagic or benthic trophic baselines.
Stable isotope TL
Pelagic
Benthic
3.0
2.3
3.0
2.4
2.8
2.2
2.9 ± 0.2
2.3 ± 0.2
Boops boops
Spicara smaris
Spicara maena
Mean
Stomach
content TL
2.5 a
3.0 a
3.2 a
2.8 ± 0.3
Labrids
Coris julis
S. tinca
S. mediterraneus
Mean
3.3 a
3.3 a
3.2 a
3.3 ± 0.1
3.3
3.3
3.2
3.3 ± 0.1
2.7
2.7
2.5
2.6 ± 0.1
Diplodus spp.
Diplodus annularis
Diplodus sargus
Diplodus vulgaris
Mean
3.4 a
3.4 a
3.1 a
3.3 ± 0.2
3.9
3.9
3.9
3.9 ± 0.4
3.3
3.2
3.3
3.3 ± 0.4
M. variegatus
Mullus surmuletus
Pagellus acarne
Pagellus erythrinus
Trigloporus lastoviza
Mean
3.4 b
3.3 a
3.7 a
3.3 a
3.5 a
3.4 ± 0.2
3.5
3.4
3.6
3.7
3.3
3.5 ± 0.2
2.8
2.8
3.0
3.0
2.7
2.9 ± 0.2
Scorpaena notata
Scorpaena porcus
Serranus cabrilla
Mean
3.5 a
4.0 a
3.4 a
3.6 ± 0.3
3.5
3.3
3.4
3.5 ± 0.1
2.8
2.7
2.7
2.8 ± 0.1
Trophic groups
Zooplankton
feeders
Muddy or sand
bottom
mesocarnivores
Macrocarnivores
Species
Benthic
piscivores
Scorpaena scrofa
4.1 a
3.4
2.8
Synodus saurus
4.5 c
3.6
2.9
a
Phycis phycis
4.1
3.6
3.0
Mean
4.2 ± 0.2
3.5 ± 0.2
2.9 ± 0.2
Pelagic
Dicentrarchus labrax
4.3d
4.6
3.9
a
piscivores
T. mediterraneus
3.5
4.9
4.2
S. viridensis
4.3e
3.7
3.1
Mean
4.0 ± 0.5
4.4 ± 0.7
3.8 ± 0.7
a: Stergiou and Karpouzi (2002) ; b: Darnaude (2005) ; c: Soares et al. (2003) ; d:
Rogdakis et al. (2010) ; e: Barreiros et al. (2002)
21
Supporting information
Supporting information may be found in the online version of the paper
Table S.1: Actual sampling at each season and each reef
Family
Species
Carangidae Trachurus mediterraneus
Centracantidae Spicara maena
Spicara smaris
Labridae Coris julis
Symphodus mediterraneus
Symphodus tinca
Moronidae Dicentrarchus labrax
Mullidae Mullus surmuletus
Phycidae Phycis phycis
Scorpaenidae Scorpaena notata
Scorpaena porcus
Scorpaena scrofa
Serranidae Serranus cabrilla
Soleidae Microchirus variegatus
Sparidae Boops boops
Diplodus annularis
Diplodus sargus
Diplodus vulgaris
Pagellus acarne
Pagellus erythrinus
Sphyraenidae Sphyraena viridensis
Synodontidae Synodus saurus
Triglidae Trigloporus lastoviza
Winter
V3
1
7
8
1
4
1
12
5
2
6
17
7
4
11
V6
5
2
1
5
2
1
3
12
2
6
4
5
1
3
33
2
5
6
5
2
1
Summer
V3
V6
3
4
1
1
2
2
3
13
6
8
10
2
3
6
9
3
2
5
2
3
7
9
1
6
4
5
1
6
1
2
1
2
22
Table S.2: Significant spatial or seasonal variations of mean isotopic ratios (13C and 15N) of
fishes on Marseilles‟ artificial reef (n ≥ 9). S: summer, W: winter. P test significance (*** for pvalue ≤ 0.001, ** for p-value ≤ 0.01, * for p-value ≤ 0.05). Post-Hoc: results of post-hoc
comparison of means
n indiv
Species
10
Spicara maena
Factor
Values (‰)
Season
S: -19.68 ± 0.27
W: -19.18 ± 0.12
28
Scorpaena porcus
Season
9
Spicara smaris
Season
35
Mullus surmuletus
Reef
S: 9.57 ± 0.35
W: 9.94 ± 0.40
S: 9.26± 0.02
W: 8.42 ± 0.78
V3: 9.83 ± 0.84
V6: 10.03 ± 0.46
test
 C
Statistics
P
Post-Hoc
7.69
*
S<W
ANOVA
9.03
**
S<W
ANOVA
9.29
*
S>W
ANCOVA
4.29
*
V3 < V6
13
ANCOVA

15N
23
Table S.3: Parameters of the linear regression between stable isotopes values and standard length of fishes. Significant relationships
are written in bold. n: number of individuals. Linear relationships were considered to be significant when n ≥ 10, r ≥ 0.5 and p ≤ 0.05.
Due to the low number (3) of individuals, no regression was performed for Phycis phycis, Synodus saurus and Trigloporus lastoviza.
δ13C
n
Carangidae
Trachurus mediterraneus
Centracanthidae Spicara maena
Spicara smaris
Labridae
Coris julis
Symphodus mediterraneus
Symphodus tinca
Moronidae
Mullidae
Scorpaenidae
Dicentrarchus labrax
Mullus surmuletus
Soleidae
Sparidae
13
-2
13
-2
13
-3
10 δ C= -21.0 + 1 x 10 SL
9
δ C= -22.3 + 2 x 10 SL
13 δ C= - 18.8 +1 x 10 SL
5
6
4
13
-3
δ C= -18.8 -4 x 10 SL
13
-3
δ C= -19.7 + 6 x 10 SL
13
-3
δ C= -20.9 + 8x 10 SL
13
-3
13
-2
35 δ C= -18.6 + 5 x 10 SL
0.32 0.37
0.75 0.00
0.61 0.03
0.06 0.82
-0.22 0.43
0.32 0.24
0.21 0.50
0.21 0.18
δ15N= 41.8 - 1 x 10-1 SL
r
p
-0.59
0.07
15
-3
-0.51
0.01
15
-3
-0.13
0.67
15
-3
δ N= 8.6 – 6 x 10 SL
δ N= 9.6 – 7 x 10 SL
δ N= 9.4 + 3x 10 SL
0.24
0.33
15
-2
0.73
0.00
15
-3
-0.45
0.09
15
-3
-0.19
0.54
15
-2
0.51
0.00
15
-3
δ N= 6.0 + 3 x 10 SL
δ N= 10.3 - 4 x 10 SL
δ N= 9.5 + 1 x 10 SL
δ N= 8.2 + 1 x 10 SL
33 δ C= -18.9 + 1 x 10 SL
0.51 0.00
δ N= 9.2 + 8 x 10 SL
0.28
0.07
Scorpaena porcus
28 δ13C= -17.1 - 3 x 10-3 SL
-0.31 0.07
δ15N= 10.1 – 3 x 10-3 SL
Serranus cabrilla
Microchirus variegatus
Boops boops
Diplodus annularis
Diplodus sargus
Diplodus vulgaris
Pagellus acarne
Sphyraenidae
δ13C= -22.4 + 1 x 10-2 SL
p
Scorpaena notata
Scorpaena scrofa
Serranidae
5
r
δ15N
Pagellus erythrinus
Sphyraena viridensis
5
13
-3
δ C= -17.1 + 5 x 10 SL
13
-5
20 δ C= -18.3 - 9 x 10 SL
7
13
-3
δ C= -18.3 + 2 x 10 SL
13
-3
33 δ C=-20.1 + 1 x 10 SL
13
-4
13
-2
48 δ C= -19.0 + 8 x 10 SL
5
δ C= -23.6 + 3 x 10 SL
13
-3
20 δ C= -17.9 - 1 x 10 SL
13
-3
13
-3
20 δ C= -18.6 + 7 x 10 SL
5
5
δ C= -19.0 + 9 x 10 SL
δ13C= -7.5 - 3 x 10-2 SL
-0.93 0.00
-0.01 0.97
0.06 0.80
0.11 0.50
0.02 0.89
0.47 0.09
-0.05 0.76
0.30 0.10
0.58 0.02
-0.65 0.04
-0.25
0.13
15
-2
0.34
0.16
15
-3
0.34
0.03
δ N= 9.7 + 2 x 10 SL
δ N= 9.0 + 5 x 10 SL
15
-3
δ N= 10.1 + 1 x 10 SL
0.08
0.74
15
-4
0.06
0.72
15
-2
δ N=8.3 + 7 x 10 SL
δ N= 5.7 + 5 x 10 SL
0.62
0.00
15
-3
-0.03
0.92
15
-3
-0.37
0.03
15
-3
-0.19
0.31
0.81
0.28
0.00
0.44
δ N= 11.7 – 1 x 10 SL
δ N= 13.1 – 1 x 10 SL
δ N= 11.2 – 4 x 10 SL
15
-2
δ N= 6.7 + 3 x 10 SL
δ15N= 7.3 – 1 x 10-2 SL
1
24